I'm trying to figure out what is likely to show up in telemetry, and how we're going to interpret the data points.
Some of the corner cases/times the TRR mode changes are:

• network changes (user connects to a different network)
• network connectivity changes (network flips)
• startup
• system suspend-resume

The patch I submitted doesn't differentiate between 0->2 and 2->0 mode switches, but I think that difference might be important.
We'd ideally want 0->2 switches to be shorter, and 2->0 switches to be longer, meaning the user spends as much time as possible in mode2.
So I'm considering making this a keyed histogram with the following keys: "0->2", "2->0", "other". An "all" bucket would also be possible, but we can infer that data based on all the keyes.

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:valentin, could you have a look please?
For more information, please visit auto_nag documentation.

Landing bug 1688076 fixed the telemetry we were seeing in respect to TRR_MODE_NOT_ENABLED being very common.
It is now down to expected levels, meaning we don't change the TRR mode more often than needed - so we don't need this telemetry.